Please note that lectures start Sept 14
***********
Schedule Fall 2026:
Module 1 (Amsterdam): Introduction to advanced statistical physics for neural networks
Tuan Pham
Lectures and exercises: Sept 14, 21, 28, Oct 5
Exam: Homework (pass/fail)
Location: Science Park D1.112 (Amsterdam)
Abstract:
Recent advances in neuroscience and AI have triggered much interest in statistical descriptions of large populations of neurons. In this set of lectures, we cover a range of tools that have been developed to understand the collective dynamics and thermodynamics of these interconnected systems, such as learning, computation, dissipation, or chaos.
We will discuss methods and topics such as: application of DMFT to study non-reciprocal networks; the interplay between plasticity and entropy production; message passing, cavity and belief propagation for sparse networks; introduction to replica and statistical field theory methods for the Hopfield model.
Prerequisites: It will be helpful if students have seen elements of stochastic processes and standard statistical mechanics.
Reference:
[1] Statistical Mechanics of Learning, A. Engel and C. Van den Broeck
[2] Statistical Mechanics of Recurrent Neural Networks I. Statics and II. Dynamics, A.C.C. Coolen
[3] Statistical Physics Methods in Optimization and Machine Learning, Florent Krzakala and Lenka Zdeborová
Module 2 (Leiden): Topological Methods (To be confirmed)
Lecturer: Luca Giomi
Lectures and exercises: Oct 12, 19, 26, Nov 2
Exam: Nov 9 (pass/fail)
Location: Leiden, Gorlaeus Building Room EM 1.21
Abstract: TBA
Module 3: TBA
Lecturer: TBA
Lectures and exercises: Nov 16, May 23, 30, Dec 7
Exam: Dec 14 (pass/fail)
Location: Utrecht
Abstract: TBA
|